1. law, govt and politics
2. legal issues
3. civil rights

Shaming the Corporation:
Social Movement Pressure and Corporate Social Responsibility
Tim Bartley
Associate Professor
Dept. of Sociology
Indiana University
bartleyt@indiana.edu
Curtis Child
Ph.D. Candidate
Dept. of Sociology
Indiana University
cchild@indiana.edu
Abstract
A growing body of work argues that social movements are key motors of institutional change. In the rise
of “corporate social responsibility” (CSR), the role of activist pressure is difficult to deny. But why would
social movement pressure matter to companies? Through what processes might “naming and shaming”
generate “sticks and stones?” We test the three most common arguments about how social
movement/stakeholder campaigns affect companies—(1) by influencing consumer behavior, (2) by
influencing investor behavior, and (3) by tarnishing valued reputations. Though these arguments have
become central to theoretical literatures on CSR and “private politics,” surprisingly little research has
tried to assess them empirically. Using data on large U.S.-based firms in the apparel and footwear
sector during the rise of the anti-sweatshop movement (1993-2000), we find little evidence for any of
these processes. Firms’ sales, market share, stock prices, and reputations in the business community
were largely unaffected by anti-sweatshop campaigns. We argue that scholars should move beyond
simplistic ideas about broad-based effects of social movement pressure to instead consider the ways in
which activist campaigns shape categories and evaluative standards in particular institutional settings. A
preliminary analysis of more focused evaluations (KLD’s “labor concern” rating) lends credence to this
alternative approach. Despite being only loosely coupled with the bottom line, activist campaigns may
lead firms to experiment with CSR practices, though we should not be surprised if such practices have
little bearing on the core production activities of firms.
Draft of Feb. 2010
Organizational scholars have recently built a sizable body of work at the intersections of social
movement and organizational theory. While some of this is an attempt to cross-fertilize theoretical
concepts (Campbell 2005; Rao, Monin and Durand 2003), much of it makes the argument that social
movements and contentious politics are key drivers of institutional change. Social movements are
argued to shape the boundaries and forms of organizations (Haveman, Rao and Paruchuri 2007; Rao
1998; Rao, Morrill and Zald 2000), generate organizational diversity (Schneiberg, King and Smith 2008),
challenge dominant frames and players in organizational fields (Fligstein 2001), and create new
industries and markets (Lounsbury, Ventresca and Hirsch 2003; Weber, Heinze and DeSoucey 2008). For
neo-institutionalist scholars of organizational behavior, social movements are increasingly replacing the
“exogenous shock” as the key driver of change. In a different strand of institutional theory, scholars
have similarly highlighted the role of pressure from activists and stakeholders in shaping organizations.
Here, movement pressure is seen as activating strategic and collective action dilemmas for firms (King,
Lenox and Barnett 2002; Spar and Yoffie 2000) and often generating new institutional arrangements for
the private provision of public goods (Abbott and Snidal 2009; Potoski and Prakash 2009). (For an
attempt to connect these two strands of work, see Bartley (2007).)
In both literatures, social movements have become prominent actors, commonly assumed to
have the power to influence firms’ behavior and potentially reconfigure the governance of industries.
Yet the general sources of this power often remain somewhat mysterious. Often, it is argued to derive
from activists’ ability to impose costs on the firm, but this is more commonly asserted than tested. In
this paper, we look for evidence to support the three most common accounts of how social movements
affect companies—one based on consumer behavior (and observable in terms of sales and market
share), one based on investor behavior (and observable in terms of stock price returns), and one based
on reputation (and observable through the Fortune corporate reputation ratings). Admittedly, our
1
approach here is one of vulgar empiricism. But given the power being imputed to social movements to
reshape industries, we feel that such an inquiry is appropriate.
We focus on the labor rights and anti-sweatshop movement of the 1990s in the U.S. and its
intersection with firms in the apparel and footwear industry. This case is appealing for several reasons.
It is often seen as a quintessential case of anti-corporate campaigns using the “name and shame”
strategy, and it influenced activism in other sectors (Seidman 2007; Spar 1998; Spar and LaMure 2003).
Furthermore, the availability of information on this sector allowed us to build a detailed dataset,
tracking the unfolding of campaigns targeting particular firms. We discuss this data and our analytical
strategy in more detail later in the paper. We begin with a brief glimpse of the importance of social
movement pressure in structuring CSR activities. We then turn to questions of how activism might
influence firms, highlight three prominent accounts, and then assess these.
CSR and Social Movements
Corporate social responsibility (CSR) for global labor standards has evolved from a few
experiments with codes of conduct and independent monitoring to a complex, layered field of standardsetting, factory auditing, and social reporting. Most major consumer products firms, from Nike to WalMart to Apple, now have some sort of policy on labor conditions in their supply chains. An entire
industry of labor compliance auditing has arisen to assess conditions at the factory level, reporting the
results to firms’ growing internal CSR staff and in some cases to associations like the Fair Labor
Association, Ethical Trading Initiative, or Social Accountability, the latter of which has certified over 1600
facilities worldwide as conforming to ethical production standards. Hundreds of firms have signed up
for the UN’s Global Compact, a voluntary initiative for CSR standard-setting and reporting, and some
have tailored their “sustainability reporting” to guidelines produced by the Global Reporting Initiative.
Though the effects of all this on labor conditions have been limited (Egels-Zanden 2007; Locke, Qin and
Brause 2007; O'Rourke 2006) and even supposed CSR leaders have faced new scandals involving child
2
labor, forced overtime, and audit fraud, even these contradictions underscore the extent to which
consumer products firms have become enmeshed in a new set of discourses, practices, and debates
about CSR.
Scholars almost universally regard the anti-sweatshop movement as the driving force behind
these changes (Bartley 2005; Esbenshade 2004; Locke, Amengual and Mangla 2009; Sabel, O'Rourke and
Fung 2000; Seidman 2007; Soule 2009). This social movement raised public awareness of global labor
rights issues through protests, media exposes, congressional hearings, and other campaign activities
through much of the 1990s. These campaigns primarily used the strategy of “naming and shaming” high
profile firms, perhaps most notably, The Gap, Nike, and Wal-Mart, each of which were nearly constant
targets.
In the apparel and footwear sector, social movement pressure played a key role in moving CSR
beyond individual codes of conduct and toward external systems for monitoring and certification
(Bartley 2007). In the midst of anti-sweatshop pressure, firms often adopted codes of conduct for their
supply chains. Yet this practice very quickly diffused to other firms, including those who had not been
singled out by activists. Consider the 39 apparel, textile, or footwear firms that are known to have had a
labor-related code of conduct at some point between 1996 and 1999.1 Most (61.5%) were targets of
activism, but more importantly, 38.5% of firms known to have codes of conduct were not targets of
activism. Social movement pressure may have jumpstarted the rise of codes of conduct, but it was far
from necessary for their adoption. (See Wetterberg (2008) for a systematic analysis of the diffusion of
labor codes of conduct, including to firms that were not targeted.)
[Table 1 here]
1
These companies were identified in inventories conducted by the U.S. Department of Labor (1996) and OECD
(1999). A later World Bank study was fully redundant with these earlier studies, at least for companies in our
industries.
3
However, when it came to CSR practices that required firms to give up even a small degree of
autonomy—such as participating in multi-stakeholder monitoring or certification associations—social
movement pressure was decisive. Among the 50 largest U.S. based apparel, footwear, and textile firms,
none participated in any of the three leading multi-stakeholder initiatives (Fair Labor Association, Social
Accountability International, or Ethnical Trading Initiative) between 2001 and 2004 without having been
targeted by activists at some point in the 1990s, as shown in Table 1. This means that social movement
pressure was a necessary condition for participation by large American firms. Though it was not a
sufficient condition, other work (Bartley 2009) has shown that higher degrees of social movement
pressure increased the odds of participation (and this effect is robust to controls for branding,
reputation, and size). Others writing about the apparel and footwear sector have similarly argued that
social movement pressure has been a prime mover and has pushed companies to engage in CSR
activities that at one point seemed unthinkable, even if they do not fully satisfy activists’ demands
(Locke, Amengual and Mangla 2009; Seidman 2007; Spar and LaMure 2003).
How do social movements affect corporations? Three common arguments
Findings such as these beg the question of how social movements get companies to do things
they would not otherwise do. Despite the explosion of work on “social movements and organizations”
(Davis et al. 2005; King and Soule 2007) and on “private politics” and corporate strategy (Baron 2003;
Baron and Diermeier 2007; Delmas and Toffel 2004; Eesley and Lenox 2006), surprisingly little work has
sought to empirically address this question. Nevertheless, three basic stories are commonly assumed or
asserted in the existing literature.
4
Consumers
One widespread idea is that activist campaigns filter into the preferences of consumers, who
then become the key forces in punishing firms viewed as irresponsible and rewarding those viewed as
responsible. Elliott and Freeman (2003) are perhaps most vocal in reducing the effect of social
movements to consumer behavior. They argue that “what gives the antisweatshop movement power is
that consumers care not only about the physical attributes of goods and services but also about the
workplace conditions associated with them” (p.28). Furthermore, “if consumers were indifferent to
those conditions, firms would ignore activist complaints about poor standards” (p.28). Although much
doubt has been cast on the accuracy and behavioral implications of such claims (Vogel 2005), the
sovereign and concerned consumer has become either an article of faith or the holy grail for much of
the existing literature.
On one hand, formal models of CSR are increasingly being built around the image of such a
consumer. The model of private provision of public goods developed by Bagnoli and Watts (2007) starts
with the assumption that consumers are sensitive (albeit differentially so) to information about the
responsibility of firms and products. Feddersen and Gilligan (2001) bring social movements into the
picture by arguing that activists are a crucial supplier of information to consumers. Here, social
movements are the conduits that allow a market for credence goods to function, since individual
consumers would often be incapable of learning about the conditions of production in global supply
chains on their own.
These specific theoretical claims are difficult to observe empirically—like many claims in
theories of CSR (McWilliams and Siegel 2001)—but other scholars have embarked on the quest to
determine how important consumer demand for fair labor conditions might be at the behavioral level.
Moving beyond the obvious problems of survey data on this topic, experiments conducted by Prasad et
al. (2004) and especially by Hiscox and Smyth (2007) have shown that some segment of consumers will
5
indeed pay more for products labeled as responsible. The idiosyncrasies of these experiments (very
unique segments of consumers, phony labels) make it difficult to estimate how large that segment might
be. Regardless, it is clear that many scholars think about the effects of social movement campaigns—
and anti-sweatshop campaigns in particular—through the lens of consumer behavior. Although most
scholars focus on consumer reactions to positive information about firms (e.g., a certification or label), it
would seem that negative information (e.g., a public scandal) might have an even greater effect, given
cognitive biases toward responding to “bad news.”2
Investors
A second possibility is that social movements may shape investment behavior and undercut
firms’ stock performance. Since investors are fewer in number, less dispersed, and more informed than
consumers, one might expect their reactions to protest and scandal to be especially powerful. Vogel
(2005) argues that the “market for virtue” among individual consumers is slight and fragile, but notes
that campaigns targeting shareholders have sometimes proven powerful.
Other work argues that even campaigns that do not explicitly target shareholders may influence
stock prices. King and Soule (2007) argue that protest activity might send signals to shareholders
regarding future market trends, stakeholder discontent, or new costs that will be imposed on the firm.
Their study of protest events from 1962-1990 finds that protest depressed firms’ stock prices by around
0.3% to 1% on average, with greater penalties found in cases where media attention was high (and
when considering larger windows of time). Notably, this penalty did not depend on whether the protest
included a call for a boycott, leading them to conclude that “boycotts do not make protests any more
effective than they already are” (p.435). Epstein and Schnietz (2002) find that 1999 protests against the
WTO in Seattle depressed the stock returns of companies in industries highlighted as abusive in their
labor or environmental practices.
2
Elliott and Freeman (p.33) find evidence of this in the survey evidence on consumers and sweatshops.
6
Reputation
Finally, many scholars have looked beyond consumer and investor behavior to argue that
activist campaigns may leverage the power of corporate reputation. As Fombrun and Shanley (1990)
put it, “just as firms compete for consumers, so also do they vie for reputational status” (p.234).
Campaigns may “name and shame” in the true sense of the word, casting a pall over the socially
constructed image of a company. Of course, shaming is only likely to be powerful when reputations are
highly salient (Braithwaite 1989). Thus, activists disproportionately target firms with high-profile, valued
reputations (Bartley and Child 2007; Lenox and Eesley 2006), seeking to turn corporate “branding” from
an asset to a liability (Klein 1999).
Some scholars see activist campaigns as significant threats to corporate reputations. Spar and
LaMure (2003) argue that even if campaigns do not have immediate effects on the bottom line, they
may threaten valued reputations to such an extent that it is in the best interest of firms to respond.
Schurman (2004) argues that dependence on corporate reputation constitutes a source of opportunity
for activists and vulnerability for firms, since “firms and industries in which brand names and/or firm
reputations are very important should be far more vulnerable to attack than industries composed of
numerous, anonymous firms producing generic products” (p.248).
Similar to arguments about consumers, notions of reputational risk are becoming foundational
in formal models of “private politics,” focused on the strategic relationships between firms and activists.
Baron and Diermeier (2007) assume that activists can harm a company (without necessarily waging a
boycott) by damaging a firm’s reputation. They use this assumption to build a broader theory of the
dynamics of private politics and the shape of activist campaigns. Indeed, theories of reputational risk
have become central to models of collective action and private regulation (King, Lenox and Barnett
2002; Potoski and Prakash 2009; see also Bartley 2007). Though a full review of the role of reputation in
7
such theories goes well beyond the scope of this paper, we do seek evidence of reputational risk in
action.
Data and Methods
Our analyses focus on a sample of the largest U.S.-based firms in the apparel, textile, and
footwear industries, including relevant retailers (i.e., specialty apparel retailers, department stores, and
discount retailers that sell apparel) (N=168). All variables used in this paper are measured yearly from
1993-2000, generating a full dataset of approximately 1100 firm-years.
To develop our sample, we used Standard & Poor’s Compustat North America Industrial Annual
File to generate a list of firms with the appropriate Standard Industrial Classification (SIC) codes and a
minimum of $300 million in either total assets or total sales at any point between 1990 and 2000.3 We
restricted the analysis to relatively large firms in order to be confident that all companies would be likely
to be covered in our data sources for measuring social movement activity. As an additional step in
constructing an appropriate sample, we used companies’ annual reports to confirm their relevance to
apparel, textile, or footwear production.4
3
To qualify for our sample, companies must be listed in Compustat during our period of analysis, issue annual
reports for shareholders, and have one of the following as their primary SIC code: textiles (SIC 22, except those
pertaining to rugs/carpets or yarn/thread), apparel manufacturing (SIC 23), footwear and accessories (SIC 302
codes pertaining to rubber footwear, SIC 31 codes pertaining to leather footwear, gloves, or handbags), wholesale
apparel or footwear (SIC 513), general/mass retailers (SIC 53), apparel retailers (SIC 56), or catalog and mail order
houses (SIC 5961). We accessed the Compustat data through Wharton Research Data Services
(wrds.wharton.upenn.edu).
4
We excluded several catalog and mail order companies that sold no apparel or footwear. Retailers were included
only if there was evidence that they were responsible for the production of some “private label” apparel, textiles,
or footwear—a common practice which was present for all but ten retailers (with the exceptions being firms like
Ross Stores, Just for Feet, Home Shopping Network, etc.). We excluded several firms that were categorized in
these SICs but which further research (using the companies’ annual reports) revealed to have only tangential
connections to apparel, textiles, or footwear during the period of our analysis (e.g. Big Lots, Claire’s Stores,
Samsonite luggage, and a company specializing in sports-related collectibles). We included one conglomerate firm
(Sara Lee), which is a major player in the apparel industry (maker of Hanes, Playtex, and other brands) but is
categorized in a different SIC.
8
To measure whether firms were targeted in anti-sweatshop campaigns, we first developed a
comprehensive database of labor-related movement activity in this industry from 1993 to 2000. While
social movement researchers often rely on newspaper reports (especially the New York Times) in order
to count instances of social movement activity (Earl et al. 2004; Oliver and Maney 2000), we utilized
more specialized sources of information to develop a rich dataset on campaigns in one particular sector.
We draw on a unique source of data on the apparel industry—the detailed coverage of industry politics
provided by the two leading trade journals—Bobbin and Women’s Wear Daily (WWD). As the industry’s
daily newspaper, WWD provided especially detailed coverage of social movement activity in and around
the industry. From these two journals, we collected 1,467 articles broadly related to the politics of labor
and coded them to generate a record of movement activity targeting particular firms. Social movement
activity was operationally defined as any attempt to force change in the labor practices of the industry
(or its firms) led by advocacy groups based outside companies, trade associations, or government. In
practice, this included protests and demonstrations at stores, leafleting and other non-confrontational
demonstrations, union solidarity pickets, and human rights lawsuits. This diversity of activity reflects
the multi-faceted character of anti-sweatshop campaigns, which often link organizing efforts at the
point of production to broader human rights or social justice concerns, involve coalitions of sponsoring
groups, and mobilize multiple forms of pressure on firms. To avoid conflating anti-corporate campaigns
with fully localized labor relations, we excluded from our purview cases that were solely about union
negotiations in the U.S. and involved only labor unions as sponsors (e.g., a strike by domestic textile
workers seeking a new contract). As a supplement the trade journal data, we consulted the leading
secondary sources on the anti-sweatshop movement for additional and confirmatory information on
companies implicated in campaigns (Armbruster-Sandoval 2005; Bonacich and Appelbaum 2000; Louie
2001; Manheim 2001; Ross 1997).5
5
The primary utility of the secondary sources was to report that a campaign was ongoing over a series of years
9
Our dataset attempts to capture the dynamics of social movement campaigns as they unfolded
in a particular sector during the rise of the anti-sweatshop movement. Comparisons to information on
anti-sweatshop activity found in the New York Times reveal that our data is more comprehensive, and
we believe it comes closer than most other studies to representing the full range of campaigns, not
merely those that received the most attention outside the industry. Contrary to images of the antisweatshop movement “starting” with attacks on Nike and the Kathie Lee Gifford scandal in 1996, our
dataset contains important but oft-forgotten earlier campaigns targeting firms like JC Penney, Liz
Claiborne, VF, and others. Later in the paper, we consider what implications the more comprehensive
nature of our dataset may have for making sense of our findings vis-à-vis research using more general
(across industries) but less detailed data on protest events.
We use the Compustat Industrial Annual database to measure firms’ performance. This includes
measures of sales, market share (sales/total sales in the firm’s 4-digit SIC), and return on assets (ROA)
(net income/total assets), all measured in real dollars, using the CPI to adjust for inflation. Our
measures of reputation come from the widely-used Fortune rating of “America’s Most Admired
Companies,” which is based on a yearly survey of executives and analysts. Because only a limited
number of firms in any particular industry segment are included in the Fortune ratings, some of our
analyses use a substantially smaller sample size. Details of our measurement strategies and additional
data sources are included in later sections of the paper.
(e.g. 1996-1998) whereas the trade journals may have neglected to report on the “middle” years of a campaign
(e.g., reporting protests in 1996 and 1998). Manheim’s (2001) database of anti-corporate campaigns, published in
the book’s appendix, proved especially useful in this regard. In general, the secondary sources overwhelmingly
confirmed what had been found in the trade journal data—listing only one corporate target that was absent from
the trade journals. Yet the trade journal data was much more detailed than the secondary sources, and revealed a
large number of corporate targets that would have been overlooked if one relied only on the secondary sources.
Our assumption is that if a company was not mentioned as a target in the trade journal articles or mentioned in
the secondary literature on the anti-sweatshop movement, then it was not a target of movement activity during
this period.
10
To control for unobserved heterogeneity among firms, we use fixed effects models whenever
possible. This is especially important since we know from other work (Bartley and Child 2007) that social
movements in this sector are more likely to target larger and higher profile firms. By using fixed effects
and including lagged dependent variables as predictors, we gear our models to examining changes over
time within companies. Furthermore, since our cases include repeated measures on the same
companies over time, we use robust standard errors (where possible). We proceed by assessing
whether anti-sweatshop campaigns had effects on (a) sales and market share, (b) stock prices, and (c)
reputation.
Do Social Movement Campaigns Affect Sales?
If indeed some consumers are as responsive to negative information about corporate abuses as
scholars often suggest they are, then one ought to find evidence of this not just in surveys or in the lab,
but in actual consumer purchases. Hiscox and Smyth (2007) frame social movement campaigns this way
in saying that “a number of publicity campaigns and consumer boycotts have been targeted to hurt the
sales of high-profile multinational firms linked to sweatshop production abroad” (p.6). Table 2 assesses
this possibility for large apparel, textile, and footwear firms from 1993-2000. Since the distribution of
sales is highly skewed, we take the natural log of this. We control for previous performance, time
trends, and for whether a company was recognized as a leading brand, since anti-sweatshop campaigns
disproportionately target heavily branded companies. Here we use a dummy variable based on whether
the firm was listed in the previous year in Advertising Age magazine’s annual list of “100 Leading
National Advertisers” (based on total advertising expenditures) or its “Marketing 100” list (highlighting
innovating “branding” campaigns and “current brand success stories” each year (Edwards 2005:70)).
[Table 2 here]
11
Model 1 examines whether being targeted in an anti-sweatshop campaign in a given year is
associated with a change in sales. The coefficient for being the target of a campaign is far from
statistical significance, providing no evidence of such an effect. In model 2, we consider whether
campaigns that received more attention (in our trade press sources) might be linked to decreased sales,
but find no evidence of this either. Model 3 sets a higher threshold—using a dummy variable to
represent firms that were targeted two years in a row (t-1 to t). Here too, the coefficient is far from
statistical significance. All of these results are substantively identical if we use data from the previous
year (t-1). Overall, contrary to assumptions that consumers would radically alter their purchases in the
wake of sweatshop allegations, we find no evidence that these campaigns shaped firms’ sales. Not
surprisingly, sales were strongly influenced by prior year’s sales, and (in some specifications) by being a
leading advertiser or innovative brand.
Even if sweatshop scrutiny did not affect firms’ absolute level of sales, it might affect their
relative market position. Models 4-6 shift the dependent variable to market share. Once again, the
coefficients for being targeted in a given year, the degree of attention a campaign garnered, and being
targeted two years in a row do not approach statistical significance. (Again, the results are substantively
identical if measures for the previous year are used instead.)
Going further, we consider whether sales declines would be more apparent only among
companies that faced the most intense sweatshop pressure. To do so, we examined the 32 firm-years in
which firms received more than the average amount of attention for sweatshop problems (defined as
more than the mean number of relevant trade journal articles per year). Only 10 of these firm-years
saw decreases in sales (in real dollars) in the midst of sweatshop scrutiny, while 22 saw increases in
sales.6 In only 12 of these 32 firm-years did market share decrease, while it increased in 20 cases. The
Gap, for instance, became a prime target of activism over labor rights abuses in El Salvador in 1995 and
6
Similar results are obtained when looking at the effect of a prior year campaign on the current year’s sales,
except that here only 9 of 32 firm-years with greater than average targeting showed sales decreases.
12
faced a new wave of legal pressure and student protest in 1999. Yet both its sales and market share
increased each year from 1995 to 2000, moving from a 20.4% market share in 1995 to a 32% share in
2000. Wal-Mart was one of the most heavily targeted firms throughout the 1990s, yet its sales and
market share increased in each of the years it was targeted. On the other hand, there are several cases
in which anti-sweatshop pressure does correspond to declines in sales or market share. Notable
examples include Guess, which had declining sales in 1997 and 1998, in the midst of multiple campaigns
over its contractors’ factories in the U.S. and Mexico, and Liz Claiborne, which saw its market share drop
from 18.4% in 1993 to 12.5% in 1995 as it faced allegations of child labor in Honduran factories.
Nevertheless, we find no evidence of a systematic link between social movement pressure and sales in
this sector.
Finally, in models 7-8, we consider whether anti-sweatshop campaigns might have conditional
effects on sales or market position, dependent on the firm’s position in the reputational hierarchy. As
others have argued, firms with solid reputations may be less vulnerable to social movement pressure
(King 2008). We therefore interact social movement targeting with a firm’s score in the Fortune
reputation ratings (relative to the mean for its industry). Neither the main effect or the multiplicative
effect are statistically significant, providing no evidence for this type of conditional influence. (It should
be noted that since not all of the firms in our sample were included in the Fortune ratings, this portion
of analysis refers only to those firms that were rated at some point.)
Do Social Movement Campaigns Affect Stock Price Returns? A Preliminary Event-Study Analysis
Event study methods have been used to examine stock market reactions to a wide variety of
events, including regulatory announcements (Binder 1985), stock buybacks (Zajac and Westphal 2004),
and celebrity endorsements (Agrawal and Kamakura 1995). Scholars have recently used this method to
estimate the effects of protest on stock prices (Epstein and Schnietz 2002; King and Soule 2007; Rock
13
2003). Event study methods use models to predict normal rates of stock price return, based on a
company’s performance in the period prior to the event of interest. This allows for a comparison of
actual returns to expected returns and the difference between them (“abnormal”) returns. Though not
without their critics, event study methods are becoming a prominent way of assessing the impacts of
activism.
Our data is of a different sort than that used in most event studies, however. While existing
studies treat anti-corporate protests as discrete events, our research shows that such protests were part
of ongoing anti-sweatshop campaigns, often lasting for years and bringing a variety of forms of pressure
to bear on particular firms. Nevertheless, we can identify some set of especially notable events in these
campaigns in order to determine whether they might affect investor activity.
To start, we generated a list of 13 discrete events (described in full in Appendix A). These
include an early expose of child labor in Wal-Mart’s supply chain in Bangladesh in 1992, the beginning of
campaign targeting The Gap over a supplier in El Salvador in 1995, the scandal involving child labor in
Honduras in the production of Kathie Lee Gifford-branded apparel sold by Kellwood and Wal-Mart in
1996, the beginning of a campaign against Kohl’s for sweatshops in Nicaragua in 2000, and others. Each
event marked an important moment in the anti-sweatshop movement, though this is not an exhaustive
list.
We performed an event study analysis using the Eventus software (accessible through Wharton
Research Data Services), which uses the relevant data from the CRSP (Center for Research in Securities
Prices) database to estimate expected returns and compare them to actual returns for a given company
and period. Following previous studies (King and Soule 2007; Zajac and Westphal 2004), we used the
simple “market model” to estimate normal returns during the event window, based on prior
performance of each company in a larger (239-day) estimation window. To minimize confounding
events, we examine relatively small event windows. We follow Zajac and Westphal in using a 2-day
14
window (from -1 to 0 days of the event) and an 11-day window (from -5 to +5) that allow for response to
early warnings or noise in the dating of events. We also use a 6 day window (from 0 to +5), which we
feel is most likely to capture immediate reactions to these events. If events occurred on non-trading
days, the windows were adjusted to begin on the next trading day.
[Table 3 here]
Table 3 shows the cumulative abnormal returns for each company as well the p-value for a test
of whether this return is statistically discernable from zero (two tailed test, based on Patell’s Z). In
general, we find little evidence that major events in the anti-sweatshop movement affected stock prices.
For only two of our events is there any evidence of negative abnormal returns. In the period following
the expose of child labor in Wal-Mart’s Bangladesh suppliers in early 1992, the company’s stock fell well
below what would be expected based on past performance. This was a highly visible and early example
of the “sweatshop stigma,” which may explain its significant effect. There is also suggestive evidence of
a similar effect when Wal-Mart became the centerpiece of a boycott of apparel made in Bangladesh in
1995, but the statistical significance of this effect depends on the event window chosen. In the vast
majority of cases, there is no compelling evidence of negative stock returns in the period around major
sweatshop scandals. In one case—Phillips Van Heusen in 1993—the onset of anti-sweatshop activism
may have coincided with higher than expected stock returns, and several other cases show returns that
are positive but not statistically significant. Note that average abnormal return in the 6-day period
following these events is negative and statistically significant. However, given the company-specific
results discussed above and the small sample size, we suspect this reflects mainly the influence of WalMart in 1992, the only statistically significant negative effect at the company level for this window.
In sum, contrary to expectations that protest and media exposes depress stock prices, we find
little evidence that major anti-sweatshop events had any consistent influence on the stock performance
of large apparel and footwear manufacturers and retailers. In supplemental analyses, we experimented
15
with shorter and longer event windows but found no additional evidence of systematic effects. While
our sample of events is not exhaustive, it includes many of the moments that most captured public
attention, including the Kathie Lee Gifford scandal, the spotlighting of Nike, and the discovery of forced
labor in a factory in El Monte, CA. (Future work will expand the set of events examined.)
Do Social Movement Campaigns Shape Corporate Reputations?
The most commonly used measure of corporate reputation comes from Fortune magazine’s
“most admired companies” ratings. Late in each year, Fortune surveys a large sample of executives and
analysts and asks them to rate the largest companies in their industry in a variety of domains, including
“quality of management,” “quality of products or services,” “wise use of corporate assets,” and others.
Previous research notes that these various attributes are highly correlated and argues for using the
overall reputation score as a measure of a firm’s global reputation in the business community (Fombrun
and Shanley 1990; Roberts and Dowling 2002).
Researchers have used these ratings to assess how reputation shapes performance (Fombrun
1996; Fombrun and Shanley 1990; Roberts and Dowling 2002), and more importantly for our purposes,
how various events shape corporate reputation. Staw and Epstein (2000), for instance, find that
linkages to popular management techniques and the implementation of TQM in the early 1990s,
improved the reputation of Fortune 100 firms, even if it did not improve their actual performance.
Flanagan and O’Shaughnessy (2005) find that layoffs in the late 1990s led to downgrading of firms’
reputation, especially for younger firms. Similarly, examining the fate of previously large firms from
1985-1994, Love and Kraatz (2009) find that firm-wide downsizing events negatively affected reputation
relative to others in their industry, though this effect diminished as downsizing became more
institutionalized and was partially mitigated by financial performance. Importantly, Love and Kraatz
view this as evidence that reputation is driven not primarily by performance or conformity to a
16
dominant logic (of shareholder value) but rather by evaluations of the organization’s character and
integrity.
By a similar rationale, one would expect anti-sweatshop campaigns to sully corporate
reputations. Sub-par labor conditions in a supply chain could be interpreted as a sign of managerial
negligence. Even if evaluators discount the claims of activists, they may expect activist “naming and
shaming” and unwanted media attention to hinder future sales, to undermine brand loyalty, or to
impose additional costs on the company relative to its competitors. Furthermore, anti-sweatshop
campaigns challenge the logic of global sourcing by arguing that lead firms ought to be held accountable
for conditions in factories owned by independent contractors. Again, even if evaluators discount
activists’ claims, they may worry that such challenges will raise the costs of production to some degree.
We examine whether anti-sweatshop pressure affected corporate reputations through fixed
effects models of apparel and footwear manufacturers and retailers from 1993-2000. Since the Fortune
ratings are only assigned for a limited number of companies in any particular industry segment, we focus
only on the 35 companies from our sample that were rated for at least two consecutive years this
period. This includes all of the largest retailers and department stores that sell apparel (e.g., Wal-Mart,
Target, Sears) as well as most of the largest apparel manufacturers (e.g., Liz Claiborne, VF, Phillips VanHeusen). A complete list of companies included in these analyses, is shown in Appendix B. We control
for common measures of financial performance—sales and return on assets (ROA)—as well as (following
Love and Kraatz 2009) changes in these factors from the previous year. Recent work trying to estimate
the extent of a financial “halo effect” on the Fortune reputation ratings finds significant effects of ROA
and sales but not other financial factors (Flanagan, O’Shaughnessy and Palmer 2008).
[Table 4 here]
In Table 4, models 1-2, the dependent variable is a firm’s relative reputation score—that is, the
difference between a firm’s reputation score and the mean for its industry. This takes note of the
17
comparative nature of the Fortune ratings (Love and Kraatz 2009) and mirror’s Roberts and Dowling’s
(2002) measurement strategy. In model 1, the coefficient for social movement targeting is in the
expected negative direction but is not statistically significant. Instead, firms’ relative reputation score
depends on their score in the previous year and their level of sales. In model 2, our measure of the
attention social movement campaigns received is not statistically significant. Substantively identical
results are obtained in models that examine the effects of targeting in the previous year (additional
models, not shown in the table). In sum, we find no convincing evidence of a systematic effect of social
movement campaigns on firm’s reputation scores. Instead, firms’ reputation in the business community
tended to be sticky over time, even when activists loudly attacked those firms. For example, Wal-Mart
remained the most admired retailer in the Fortune survey throughout the 1990s, even as it was
repeatedly facing the sweatshop stigma.
Models 3-4 shift to a dichotomous dependent variable, in order to assess the common claim
that social movement pressure spoils the good reputations of companies that have heavily invested in
reputational capital. We constructed a dummy variable where 1 represents companies that move from
having above average reputation ratings (relative to their industry groups) to having below average
ratings or being dropped from the Fortune ratings altogether. Firms with stable or improving
reputations are coded as 0. Models 3-4 show no evidence of a spoiling effect.7 The coefficients for
social movement targeting and for the extent of attention it received are both non-significant.
(Furthermore, if they were significant, negative coefficients in this model would mean that being
7
Fixed effects are not applicable in these logistic regression models, due to insufficient variation on the dependent
variable, which would cause some companies to be dropped. We instead report the results of random effects
models, which allow those companies to be included, while still helping to isolate changes over time.
Unfortunately, these do not allow for robust standard errors. We therefore ran additional population-averaged
models, which do allow for robust standard errors in a logistic regression framework. These generated
substantively identical results to those reported.
18
targeted decreased the chances of having a declining reputation.)8 Overall, despite the prominence of
“naming and shaming” strategies, we find no evidence that this shaming hurt firms’ reputations within
the business community.
[Table 5 here]
Finally, it is worth considering the possibility that it is the initial onset of an anti-sweatshop
campaign that might shape corporate reputation. Initially, sweatshop allegations might generate
uncertainty in the business community, even if this diminished over time, as companies developed more
sophisticated responses. In Table 5, we examine only the first instance of targeting, dropping companies
from the analysis once they have experienced this event. Here, we do find evidence that the initial
onset of a social movement campaign was associated with a statistically significant decline in firms’
relative reputation scores. To unpack this further, model 2 considers whether the onset of a campaign
might matter differently for firms that were already at the top of reputational hierarchies versus those
who were not. We include a dummy variable for having a reputation score above the industry average
in the previous year and interact this with our measure of social movement targeting. The significant
negative main effect in this model indicates that it was only firms who had below average prior
reputations that experienced a penalty as a result of anti-sweatshop targeting. The multiplicative effect
is positive, though not statistically significant. We take this pattern of findings as suggestive but not
definitive evidence that the onset of social movement activism mattered for the reputation of some
firms (i.e., those in low-status positions to begin with) but not for all. Declining degrees of freedom
make it difficult to establish this finding more definitively. Importantly, this is the only evidence we have
found of activism having any negative effect on reputation. Our overall conclusion is that antisweatshop campaigns failed to penetrate reputations within the business community, though they may
8
We tried a variety of other specifications of the dependent variable, including one based on intra-industry
rankings (following Love and Kraatz 2009) and one based on the absolute reputation score. No evidence of a social
movement effect could be found, regardless of how we defined the dependent variable.
19
have had this effect under some particular (but rare) circumstances.9 The initiation of an antisweatshop campaign may shape some companies’ relative positions in reputational hierarchies, but
such campaigns are far from sufficient to affect the overall status ordering of firms or to sully good
reputations.
What’s Left? Toward Alternative Conceptions of How Social Movements Affect Firms
The lack of systematic evidence for effects of anti-sweatshop campaigns is striking. Analysts
often point to this sector for examples of how movements have influenced their corporate targets, and
scholars often suggest that social movement influence on firms travels through consumers, investors, or
reputational investments. Of course, we have not proven that such effects are non-existent, only that
they are exceedingly difficult to find when one looks at the industry in general. In this section, we
consider some possible reasons for these null findings and sketch some alternative ways forward.
One possible reason for these null findings is that the more comprehensive one’s data is
(meaning, the more it captures a wide swath of campaigns, even if in a single sector) the harder it
becomes to capture social movement effects. Most social movement campaigns, including those
targeting firms, generate little more than a few ripples. Only some campaigns generate enough
attention to be picked up in national media sources; and even then, the high profile events that are
reported have often been preceded by years of campaign work, smaller events, and failed campaigns
involving other companies. By relying on generalist national media sources for their data, other scholars
may actually be measuring the effects of a relatively successful subset of campaigns.
9
Extending King (2008), we also ran models to assess whether campaigns might harm poor performing firms but
not others, using an interaction between being targeted and ROA. The sign of the coefficients are consistent with
this possibility, but they do not reach conventional levels of statistical significance. We also ran models to see if
campaigns involving especially emotional issues (child labor) or those sponsored by particular kinds of activist
groups had different effects. We found no evidence of such differences.
20
Even so, our null findings also reflect limitations in the operationalization of arguments about
consumer demand and corporate reputation. Movements may influence the reputation of firms among
some audiences, just not among the executives and analysts whose opinions make up the standard
measure of corporate reputation—the Fortune reputation ratings. The lack of a measure of reputation
among consumers or potential employees is obviously a limitation of this and other studies. But it also
points out the need to think more carefully about reputation as a construct. The reputation of WalMart, for instance, varies dramatically depending on who one asks. Furthermore, our evidence does not
reject the overall notion that consumers utilize activist claims in their purchases, but it does demand
that scholars better specify the locations in which that action is most likely to occur.
Our findings could also be read as indicating that anti-sweatshop campaigns do not actually
matter for firms. We think this conclusion would be premature. Rather than either rejecting or
resuscitating ideas about systematic effects of social movement campaigns (among consumers,
investors, etc. writ large), we instead consider how the effects of movements may be quite focused in
particular institutional locations. We briefly sketch three promising ways forward.
First, it may be that social movements are enforced through markets, but this occurs mainly in
settings in which organized actors enjoy market power or control over access. Seidman (2007), for
instance, traces the market power of anti-sweatshop campaigns mainly to universities and institutional
investors, rather to consumers and investors writ large. Work by Wetterberg (2008), extending Fligstein
(1996), seeks to develop a broader account of how specific market access points have proven powerful
for the anti-sweatshop movement. This move, from generalized consumers, to organized market power
(“islands of conscious power in this ocean of unconscious cooperation,” as the famous line from Coase
goes), is one step toward more accurate accounts of how social movements influence the bottom line.
Second, even if social movements inspire only a small degree of behavioral change among
consumers and investors, this may disrupt firms’ resource environments enough to lead them to gamble
21
on CSR or other new practices. More important than measuring the extent of consumer (or investor)
demand, therefore, would be assessing how much evidence of consumer demand it takes for firms to
respond and how they monitor and assess the cost effectiveness of new practices over time. To be sure,
scholars who emphasize consumer and investor enforcement of social movements share with us a sense
that in competitive settings, relatively small changes in preferences may be enough to lead companies
to experiment with new practices. But to our knowledge, researchers have not sought to uncover how
this plays out, either in terms of market signals or in terms of the relevant organizational decisionmaking processes.
Finally, social movements may shape how actors understand markets not at the level of the
general population (as reputation arguments suggest) but by inspiring and shaping new systems of
evaluation and categorization. The information provided through activism may inform not only
individual consumer choices (ala Feddersen and Gilligan (2001)), but perhaps more powerfully, the
findings of specialized evaluators and raters in the organizational field. Social movements are most
likely to shape newer rating bodies, often operating at the fringes of markets, not in their mainstream.
But such systems may get institutionalized in ways that carry broader effects down the line. Here, we
echo organizational sociologists who have shown how rating and categorization systems serve as
mediators of institutional change (Rao, Monin and Durand 2005) and enforce particular logics of
appropriateness (Espeland and Sauder 2006; Zuckerman 1999).
To illustrate this third path forward, we explore the linkages between anti-sweatshop campaigns
and the specialized ratings of companies done by KLD Research and Analytics, a leading source of
information for the socially responsible investment community and “the most exhaustive and widely
used measure of CSR” (Vogel 2005:31). While scholars often view KLD as an “objective” measure of CSR
performance, we view the KLD ratings as socially constructed system of categorization and rating, and
22
we join those who are beginning to unpack the contextualized meanings of these ratings (Chatterji,
Levine and Toffel 2009).
We argue that KLD ratings are far from arbitrary, but they are based on information that
becomes available largely because of prior social movement campaigns. Activist campaigns have a
unique capacity to generate new information about markets, particularly by utilizing transnational
networks and mobilizing knowledge and information that are often hidden or suppressed. Emerging out
of the Domini 400 investment index, KLD began in the 1990s to generate binary indicators of specific
“strengths” of and “concerns” about firms. While most of these focus on community involvement,
corporate governance, environmental performance, and employee relations, in the late 1990s, KLD
began looking at “labor rights concerns” pertaining to firms’ supply chains.10 This new rating in itself can
be seen as a result of the anti-sweatshop movement, which fought to make global labor conditions
relevant to firms and their audiences. But beyond merely generating a niche for a firm like KLD to fill,
social movement campaigns may shape the specific ratings of firms.
[Table 6 here]
The models in Table 6 take the existence of a KLD “labor rights concern” as the dependent
variable (looking at all apparel, textile, and footwear firms that were rated by KLD). Model 1 is crosssectional and uses logistic regression to examine whether firms that were targeted at some point in the
1990s (1993-2000) were more likely than others to receive a labor rights concern rating sometime
between 1999 and 2003. Indeed, those firms that faced social movement campaigns were far more
likely than their competitors to have their labor rights situation listed as a concern by KLD. Models 2 and
3 shift to longitudinal analyses to examine whether recent social movement activity gets translated into
subsequent KLD ratings. Model 2 shows that being targeted by activists in the previous three years
10
See “2007 Environmental, Social and Governance Ratings Criteria,” KLD Research & Analytics, Inc. Downloaded
14 January 2008 from www.kld.com. KLD later created a “labor rights strength” indicator, but only a few firms
have yet received this rating.
23
significantly increases the likelihood of receiving this negative attention from KLD in a given year.
Similarly, in model 3 we find that the amount of attention to a firm-specific anti-sweatshop campaign in
the previous three years is also associated with a greater risk of negative attention from KLD. The
control for firm size is not statistically significant, but the positive and significant effect of prior labor
concern ratings suggests that KLD’s evaluations are somewhat sticky over time.11
These results suggest that social movement campaigns were reflected in KLD’s focused
evaluations of firms. Of course, we cannot definitively prove that activists and KLD did not
independently converge on the same judgments about particular firms. But given that there is no
straightforward way to discover “who is really using sweatshops” across an entire industry—in part
because such production methods are often endemic rather than exceptional, and in part due to the
dispersed nature of global supply chains—it is reasonable to expect that KLD would follow the lead of
social movements in identifying “labor rights concerns.” More generally, in situations in which social
movement campaigns bring previously hidden information to light and in which the costs of gathering
such information independently are high, we would expect specialized evaluators to rely on information
generated by social movements. Given the lack of evidence that anti-sweatshop pressure affected sales,
11
A closer look at particular cases illustrates the linkage between social movement campaigns and the KLD ratings.
Take the example of JC Penney, which though not as commonly associated with sweatshops in public discourse as
are Wal-Mart or Nike, faced sustained anti-sweatshop pressure throughout the 1990s, especially over its labor
practices in Central America and the Caribbean. Initially, the retailer was targeted for carrying clothes made by
Philips Van Heusen workers in Guatemala, where trade unions were struggling to gain bargaining rights and escape
repression, and by the Leslie Fay company, which faced a long battle with American unions as it sought to
outsource more of its production to offshore sweatshops. Then in 1995, the National Labor Committee (NLC)
began a sustained attack against the retailer, highlighting workplace abuses and poverty-level wages in supplier
factories in Haiti and Nicaragua, the latter of which was exposed by the television news tabloid Hard Copy. In
1999, JC Penney was among a number of firms charged in a class action lawsuit with complicity in the violation of
U.S. labor laws in the territory of Saipan. That same year, it was among the first set of apparel firms to be listed as
a “labor rights concern” by KLD and continued to receive this rating in the following years. More broadly, the data
suggests that the pattern of social movement targeting is indeed reflected in the KLD ratings. 35 apparel, textile,
or footwear firms were rated as a labor rights concern sometime between 1999 and 2003. 60% of these had been
the target of social movement activism between 1993 and 1999 (and other evidence suggests that several others
were targets after the period covered by our social movement data, which ends in 2000). Twelve other apparel,
textile, and footwear firms were rated by KLD during this time but were not identified as a “labor rights concern.”
Of these, the vast majority (75%) had not been targeted by the anti-sweatshop movement.
24
such focused evaluations may be among the most important drivers of companies’ interest in CSR—as
well as rationalized symbols of CSR themselves.
To be sure, this paper has not exhausted the possible processes through which social
movements may influence firm behavior. Processes involving firm “insiders,” alternative organizational
forms, and other types of reputational processes are worthy of further consideration. Our findings
highlight the importance of delving further into the routes by which social movements that rely on
“naming and shaming” may produce “sticks and stones” and other things that shape the actions of
firms. On one hand, the routes of influence may be more varied than most discussions suggest. On the
other hand, they are also more partial and fragmentary than much of the discourse about CSR would
have it seem. One should therefore not be surprised if firms’ engagement with CSR is widespread but
somewhat shallow.
25
Table 1. Social movement pressure as a necessary condition for participation in multi-stakeholder
initiatives (MSIs), 50 largest U.S. apparel, textile, and footwear firms
Participation in MSI?
No
Yes
Total
Targeted by activists in 1990s
No
Yes
22 (100%)
19 (68%)
0 (0%)
9 (32%)
22
28
26
Total
41 (82%)
9 (18%)
50
Table 2. Testing the effect of social movement pressure on sales and market share. Fixed effects models.
VARIABLES
log sales (t-1)
Targeted by anti-sweat campaign (t)
(1)
log sales
0.8568***
(0.0198)
-0.0134
(0.0181)
Extent of attention to campaign (t)
(2)
log sales
0.8576***
(0.0200)
1994
1995
1996
1997
1998
1999
2000
(4)
marketshare
0.6858***
(0.0278)
0.0007
(0.2945)
-0.0026
(0.0021)
Targeted two years in a row (t-1 to t)
Leading advertiser/brand (t-1)
(3)
log sales
0.8102***
(0.0314)
0.0340
(0.0228)
0.0252**
(0.0123)
0.0203
(0.0184)
-0.0305
(0.0248)
0.0023
(0.0251)
0.0243
(0.0215)
0.0078
(0.0197)
0.0040
(0.0231)
0.0344
(0.0229)
0.0256**
(0.0123)
0.0200
(0.0184)
-0.0308
(0.0246)
0.0024
(0.0248)
0.0230
(0.0214)
0.0068
(0.0194)
0.0038
(0.0229)
1.0542***
(0.1330)
1098
168
0.795
1.0483***
(0.1339)
1103
168
0.795
(5)
marketshare
0.6876***
(0.0279)
(6)
marketshare
0.5741***
(0.0438)
0.0042
(0.0317)
0.0464*
(0.0277)
-0.0042
(0.0173)
-0.0527**
(0.0230)
-0.0202
(0.0266)
0.0047
(0.0222)
-0.0089
(0.0220)
-0.0096
(0.0270)
0.8493**
(0.4101)
-0.1688
(0.2176)
-0.8023*
(0.4488)
-0.5496*
(0.2984)
-0.2266
(0.2356)
-0.1853
(0.3019)
-0.0323
(0.3038)
0.0735
(0.3491)
0.8439**
(0.4088)
-0.1695
(0.2177)
-0.8029*
(0.4480)
-0.5491*
(0.2953)
-0.1898
(0.2360)
-0.1705
(0.3002)
-0.0670
(0.3010)
0.0478
(0.3479)
1.4010***
(0.2131)
956
166
0.753
3.4725***
(0.4051)
1099
168
0.552
3.4557***
(0.4097)
1104
168
0.551
0.2172
(0.3878)
1.2422**
(0.4888)
-0.6314
(0.4352)
-0.4405
(0.3159)
-0.1724
(0.3003)
-0.1016
(0.3188)
0.0927
(0.3089)
0.2136
(0.3937)
Targeted * Rep. score
N (firm-years)
N (firms)
R-squared (within)
Robust standard errors in parentheses
(8)
marketshare
0.6098***
(0.1071)
0.2071
(0.7378)
0.0590**
(0.0288)
0.0503**
(0.0207)
0.0524
(0.0391)
0.0409*
(0.0213)
0.0627**
(0.0234)
0.0333
(0.0276)
0.0307
(0.0297)
0.0420
(0.0276)
0.0526***
(0.0155)
0.0081
(0.0236)
1.5021***
(0.5230)
216
40
0.721
0.7428*
(0.4370)
-0.2107
(0.4933)
-0.8577
(0.9740)
-0.3567
(0.6540)
0.0465
(0.7708)
-1.0548
(0.8747)
-0.5902
(0.9410)
0.2687
(1.4109)
0.1003
(0.5734)
-0.1916
(0.6590)
10.1893***
(2.8377)
216
40
0.406
0.0012
(0.0186)
Reputation score (t)
Constant
(7)
log sales
0.8257***
(0.0603)
-0.0248
(0.0234)
2-tailed tests: *** p<0.01, ** p<0.05, * p<0.1
4.4884***
(0.6180)
957
166
0.433
Table 3. Results of Event Study of Abnormal Stock Returns
Company and Date
Wal-Mart, Jan 1. 1992
Phillips Van-Heusen, May 24 1993
Leslie Fay, May 31 1994
Liz Claiborne, Sept. 22 1994
Wal-Mart, April 25 1995
Gap, July 21 1995
Target/Dayton Hudson, Aug. 9 1995
Wal-Mart, April 30 1996
Kellwood, April 30 1996
Nike, July 15 1996
Nike, April 20 1998
Gap, March 6 1999
Kohl’s, July 15 2000
Mean for all events
Abnormal return,
2 day period
(-1 to 0 days)
-3.30%
(.1351)
5.04%*
(.0482)
3.71%
(.5614)
4.24%
(.1532)
-3.59%
(.1173)
2.89%
(.3410)
-1.05%
(.6025)
0.24%
(.9259)
-1.82%
(.6141)
-0.52%
(.8823)
.12%
(.9703)
0.9%
(.8252)
0.4%
(.9189)
0.56%
(.7482)
Abnormal return,
11-day period
(-5 to +5 days)
-13.87%**
(.0072)
11.03%+
(.0651)
15.22%
(.3089)
6.79%
(.3298)
-10.73%*
(.0457)
3.38%
(.6466)
-3.60%
(.4452)
2.44%
(.6821)
0.92%
(.9143)
-4.86%
(.4638)
9.60%
(.2181)
2.72%
(.7773)
2.01%
(.8248)
1.62%
(.9287)
In parentheses: P-values for 2-tailed tests, based on Patell’s Z
+ p<.10 * p<.05 ** p<.01 *** p<.001
28
Abnormal return,
6-day period
(0 to +5 days)
-12.45%**
(.0011)
3.56%
(.4204)
-7.5%
(.4975)
-1.81%
(.7247)
-6.16%
(.1204)
-7.2%
(.1674)
-4.3%
(.2165)
0.7%
(.8714)
5.6%
(.3719)
-6.4%
(.1518)
10.88%+
(.0585)
-3.33%
(.6391)
1.15%
(.8635)
-2.1%+
(.0739)
Table 4. Testing the effect of social movement pressure on corporate reputation ratings. Fixed and random effects models
VARIABLES
prior reputation (score at t-1)
Targeted by anti-sweat campaign (t)
(1)
Rel. rep score
(2)
Rel. rep score
0.4982***
(0.0757)
-0.0079
(0.0546)
0.5034***
(0.0720)
Extent of attention to campaign (t)
Return on assets (ROA) (t)
Diff. in ROA (t-1 to 1)
Sales (t)
% change in sales (t-1 to t)
Constant
N (firm-years)
N (firms)
R-squared (within)
1.6646
(1.1330)
0.4835
(0.7310)
0.0096***
(0.0012)
0.3468
(0.2760)
-0.2253***
(0.0507)
0.0141
(0.0208)
1.7277
(1.1231)
0.5163
(0.7336)
0.0094***
(0.0012)
0.3684
(0.2773)
-0.2437***
(0.0520)
192
35
0.396
192
35
0.399
(3)
Drop from
positive rep.
-1.2562*
(0.6421)
-0.2832
(0.6288)
(4)
Drop from
positive rep.
-1.2162*
(0.6367)
-11.5826
(7.0700)
-6.1968
(5.0891)
-0.0198
(0.0207)
-1.4529
(2.4669)
7.9214*
(4.2252)
-0.0841
(0.1863)
-11.4985
(7.1114)
-6.5239
(5.2839)
-0.0207
(0.0203)
-1.2648
(2.4482)
7.6287*
(4.1890)
127
31
127
31
Standard errors in parentheses (robust standard errors for models 1-2)
2-tailed tests: *** p<0.01, ** p<0.05, * p<0.1
Models 1-2 use fixed effects. Models 3-4 are logistic regression models with random effects. (Additional analyses using population-averaged
logit models with robust standard errors produce substantively identical results.)
29
Table 5. Effects of first targeting on corporate reputation ratings. Fixed effects models
(1)
Rel. rep score
0.2192**
(0.1022)
-0.2469**
(0.1047)
1.0686
(0.6882)
0.0383
(0.6547)
-0.0069
(0.0196)
0.4683
(0.3947)
VARIABLES
prior reputation (score at t-1)
Targeted by anti-sweat campaign (t)
Return on assets (ROA) (t)
Diff. in ROA (t-1 to 1)
Sales (t)
% change in sales (t-1 to t)
0.1329
(0.2044)
(2)
Rel. rep score
0.2001
(0.1420)
-0.3308*
(0.1774)
1.0134
(0.6827)
0.1306
(0.6355)
-0.0093
(0.0202)
0.4732
(0.4181)
0.0385
(0.1269)
0.1703
(0.2332)
0.1374
(0.2163)
112
31
0.197
112
31
0.203
Leading reputation (t-1)
Targeted * Leading reputation
Constant
N (firm-years)
N (firms)
R-squared (within)
Robust standard errors in parentheses
2-tailed tests: *** p<0.01, ** p<0.05, * p<0.1
30
Table 6. Social movement campaigns reflected in KLD “labor rights concern” ratings
DV: KLD “labor rights concern” rating (0/1)
(1)
Any LR concern,
1998-2003
(cross-sectional)
Past KLD “labor concern” assessment
Ever targeted by anti-sweat campaign (‘93-‘00)
(2)
LR concern,
1998-2001
(longitudinal)
3.50***
(0.41)
2.11*
(0.97)
Targeted by anti-sweat campaign in past 3 years
3.56**
(1.20)
Extent of attention to campaign (# articles in past 3 years)
Firm size (log assets)
Constant
N (firms)
N (firm-years)
Wald Chi-squared test
Likelihood Ratio Chi-squared test
(3)
LR concern,
1998-2001
(longitudinal)
2.77***
(.515)
-0.40
(0.34)
3.29
(2.46)
-0.22
(0.35)
-0.17
(2.51)
.788*
(.456)
-.066
(.314)
-.919
(2.295)
47
34
93
199.01***
34
93
78.02***
5.98+
Note: The risk set consists of those apparel, textile, and footwear firms evaluated by KLD. Dependent variable is equal to 1 if the firm received a “labor
concern” assessment for the year.
SEs in parentheses
+ p < 0.10 * p < 0.05 ** p < 0.01 *** p < 0.001
(one-tailed test for social movement effects; two-tailed otherwise)
31
Appendix A: Companies and events included in the analysis of Cumulative Abnormal Returns
Wal-Mart, Jan 1. 1992: Immediately following Dateline NBC expose of Bangladeshi child laborers
Phillips Van-Heusen, May 24, 1993: Beginning of national wave of protests over factory in Guatemala
Leslie Fay, May 31, 1994: Apparel workers union first calls for boycott of Leslie Fay
Liz Claiborne, Sept. 22 1994: Senate testimony by a child laborer from Liz Claiborne contractor in
Honduras
Wal-Mart, April 25, 1995: Beginning of a larger Bangladesh apparel boycott campaign, focusing largely
on Wal-Mart
Gap, July 21, 1995: Beginning of the campaign over the Mandarin factory in El Salvador
Target/Dayton Hudson, Aug. 9, 1995: Subsidiary (Mervyn’s) is linked to slaveshop in El Monte, CA (the
first linkage to a major retailer)
Wal-Mart, April 30, 1996: Initial expose of child labor in Kathie Lee Gifford line, sold exclusively at WalMart
Kellwood, April 30, 1996: Initial expose of child labor in Kathie Lee Gifford line, sold at Wal-Mart but
sourced through Kellwood
Nike, July 15, 1996: Beginning of a major wave of store protests
Nike, April 20 1998: Lawsuit filed charging Nike with lying about its factory conditions; campus antisweatshop movement also targeting Nike at this time.
Gap, March 6 1999: Major wave of student protests at Gap stores in various parts of the country,
partially in response to recent lawsuit over labor abuses in Saipan
Kohl’s, July 15, 2000: Beginning of a campaign over a factory in Nicaragua
32
Appendix B. Firms included in the analysis of reputation ratings
Burlington Industries
Target/Dayton Hudson
Dillards
Federated Dept. Stores
Fieldcrest Cannon
The Gap
Hartmarx
Kmart/Sears Holdings
Kellwood
Limited Brands
Liz Claiborne
May Dept. Stores
Nike
Nordstrom
JC Penney
Reebok
Russell
Sara Lee
Sears Roebuck
Springs Industries
UNIFI
VF
TJX
Wal-Mart
Leslie Fay
Crystal Brands
Fruit of the Loom
Saks
Warnaco
Levi Strauss
Cone Mills
Jones Apparel
Westpoint Stevens
Costco
Polo Ralph Lauren
Note: This constitutes the set of firms in our sample of large apparel, footwear, or textile manufacturers
or retailers that were rated in at least two consecutive years in the Fortune reputation ratings (thus
allowing for analysis of change over time).
33
References
Abbott, Kenneth, and Duncan Snidal. 2009. "Strengthening International Regulation Through
Transnational New Governance: Overcoming the Orchestration Deficit." Vanderbilt Journal of
Transnational Law 42.
Agrawal, Jagdish, and Wagner A. Kamakura. 1995. "The Economic Worth of Celebrity Endorsers: An
Event Study Analysis." The Journal of Marketing 59:56-62.
Armbruster-Sandoval, Ralph. 2005. Globalization and Cross-Border Labor Solidarity in the Americas: The
Anti-Sweatshop Movement and the Struggle for Social Justice. New York: Routledge.
Bagnoli, Mark, and Susan G. Watts. 2007. "Selling to Socially Responsible Consumers: Competition and
the Private Provision of Public Goods." Journal of Economics & Management Strategy 12:419445.
Baron, David P. 2003. "Private Politics." Journal of Economics & Management Strategy 12:31-66.
Baron, David P., and Daniel Diermeier. 2007. "Strategic Activism and Nonmarket Strategy." Journal of
Economics & Management Strategy 16:599-634.
Bartley, Tim. 2005. "Corporate Accountability and the Privatization of Labor Standards: Struggles over
Codes of Conduct in the Apparel Industry." Research in Political Sociology 12:211-244.
—. 2007. "Institutional Emergence in an Era of Globalization: The Rise of Transnational Private
Regulation of Labor and Environmental Conditions." American Journal of Sociology 113:297-351.
—. 2009. "Standards for Sweatshops: The Power and Limits of Club Theory for Explaining Voluntary
Labor Standards Programs." in Voluntary Programs: A Club Theory Perspective, edited by Aseem
Prakash and Matthew Potoski. Cambridge, MA: MIT Press.
Bartley, Tim, and Curtis Child. 2007. "Shaming the Corporation: Reputation, Globalization and the
Dynamics of Anti-Corporate Movements." Working paper, Department of Sociology, Indiana
University.
Binder, John J. 1985. "Measuring the Effects of Regulation with Stock Price Data." The RAND Journal of
Economics 16:167-183.
Bonacich, Edna, and Richard P. Appelbaum. 2000. Behind the Label: Inequality in the Los Angeles Apparel
Industry. Berkeley: University of California Press.
Braithwaite, John. 1989. Crime, Shame, and Reintegration. New York: Cambridge University Press.
Campbell, John L. 2005. "Where Do We Stand? Common Mechanisms in Organizations and Social
Movements Research." Pp. 41-68 in Social Movements and Organization Theory, edited by
Gerald F. Davis, Doug McAdam, W. Richard Scott, and Mayer N. Zald. New York: Cambridge
University Press.
Chatterji, Aaron K., David I. Levine, and Michael W. Toffel. 2009. "How Well Do Social Ratings Actually
Measure Corporate Social Responsibility?" Journal of Economics & Management Strategy
18:125-169.
Davis, Gerald F., Doug McAdam, W. Richard Scott, and Mayer N. Zald (Eds.). 2005. Social Movements
and Organization Theory. New York: Cambridge University Press.
Delmas, Magali, and Michael W. Toffel. 2004. "Stakeholders and environmental management practices:
an institutional framework." Business Strategy and the Environment 13:209-222.
Earl, Jennifer, Andrew Martin, John D. McCarthy, and Sarah A. Soule. 2004. "The Use of Newspaper Data
in the Study of Collective Action." Annual Review of Sociology 30:65-80.
Edwards, Larry. 2005. "Starting with a name only: Connecting marketing to people, one of `Ad Age's'
favorite projects is born." Advertising Age March 28, 2005, p.70.
Eesley, Charles E., and Michael J. Lenox. 2006. "Firm Responses to Secondary Stakeholder Action."
Strategic Management Journal 27:765-781.
34
Egels-Zanden, Niklas. 2007. "Suppliers' Compliance with MNCs' Codes of Conduct: Behind the Scenes at
Chinese Toy Suppliers." Journal of Business Ethics 74.
Elliott, Kimberly Ann, and Richard B. Freeman. 2003. Can Labor Standards Improve Under Globalization?
Washington, DC: Institute for International Economics.
Epstein, Marc J., and Karen E. Schnietz. 2002. "Measuring the Cost of Environmental and Labor Protests
to Globalization: An Event Study of the Failed 1999 Seattle WTO Talks." International Trade
Journal 16:129-160.
Esbenshade, Jill. 2004. Monitoring Sweatshops: Workers, Consumers, and the Global Apparel Industry.
Philadelphia: Temple University Press.
Espeland, Wendy Nelson, and Michael Sauder. 2006. "Rankings and Reactivity: How Public Measures
Recreate Social Worlds." American Journal of Sociology 113:1-40.
Feddersen, Timothy J., and Thomas W. Gilligan. 2001. "Saints and Markets: Activists and the Supply of
Credence Goods." Journal of Economics & Management Strategy 10:149-171.
Flanagan, David J., and K.C. O’Shaughnessy. 2005. "The Effect of Layoffs on Firm Reputation." Journal of
Management 31:445-463.
Flanagan, David J., K.C. O’Shaughnessy, and Timothy B. Palmer. 2008. "Re-examining the Financial
Performance Halo in the Fortune Reputation Survey." Paper presented at the Midwest Academy
of Management conference,
http://www.midwestacademy.org/Proceedings/2008/papers/Flanagan%20etal_51.pdf.
Fligstein, Neil. 1985. "The Spread of the Multidivisional Form Among Large Firms, 1919-1979." American
Sociological Review 50:377-391.
—. 1996. "Markets as Politics: A Political-Cultural Approach to Market Institutions." American
Sociological Review 61:656-73.
—. 2001. "Social Skill and the Theory of Fields." Sociological Theory 19:105-25.
Fombrun, Charles J. 1996. Reputation: Realizing Value from the Corporate Image. Boston: Harvard
Business School Press.
Fombrun, Charles, and Mark Shanley. 1990. "What's in a Name? Reputation Building and Corporate
Strategy." Academy of Management Journal 33:233-258.
Haveman, H. A., H. Rao, and S. Paruchuri. 2007. "The winds of change: The progressive movement and
the bureaucratization of thrift." American Sociological Review 72:117-142.
Hiscox, Michael J., and Nicholas F. B. Smyth. 2007. "Is There Consumer Demand for Improved Labor
Standards? Evidence from Field Experiments in Social Labeling." Working paper, Dept. of
Government, Harvard University.
King, Andrew A., Michael J. Lenox, and Michael L. Barnett. 2002. "Strategic Responses to the Reputation
Commons Problem." Pp. 393-406 in Organizations, Policy, and the Natural Environment, edited
by Andrew J. Hoffman and Marc J. Ventresca. Stanford, CA: Stanford University Press.
King, Brayden G. 2008. "A Political Mediation Model of Corporate Response to Social Movement
Activism." Administrative Science Quarterly 53:395-421.
King, Brayden G., and Sarah A. Soule. 2007. "Social Movements as Extra-institutional Entrepreneurs: The
Effect of Protest on Stock Price Returns." Administrative Science Quarterly 52:413-442
Klein, Naomi. 1999. No Logo: Taking Aim at the Brand Bullies. New York: Picador.
Lenox, Michael J., and Charles E. Eesley. 2006. "Private Environmental Activism and the Selection and
Response of Firm Targets." Working paper, Fuqua School of Business, Duke University.
Locke, Richard, Matthew Amengual, and Akshay Mangla. 2009. "Virtue out of Necessity?: Compliance,
Commitment and the Improvement of Labor Conditions in Global Supply Chains." Politics &
Society 37:319-351.
Locke, Richard, Fei Qin, and Alberto Brause. 2007. "Does Monitoring Improve Labor Standards? Lessons
from Nike." Industrial & Labor Relations Review 61:3-31.
35
Louie, Miriam Ching Yoon. 2001. Sweatshop Warriors: Immigrant Women Take on the Global Factory.
Cambridge, MA: South End Press.
Lounsbury, Michael, Marc J. Ventresca, and Paul M. Hirsch. 2003. "Social Movements, Field Frames and
Industry Emergence: A Cultural-Political Perspective on U.S. Recycling." Socio-Economic Review
1:71-104.
Love, E. Geoffrey, and Matthew Kraatz. 2009. "Character, Conformity, or the Bottom Line? How and
Why Downsizing Affected Corporate Reputation." Academy of Management Journal 52:314-335.
Manheim, Jarol. 2001. The Death of a Thousand Cuts: Corporate Campaigns and the Attack on the
Corporation. Mahwah, NJ: Lawrence Erlbaum Associates.
McWilliams, A., and D. Siegel. 2001. "Corporate social responsibility: a theory of the firm perspective."
Academy of Management Review 26:117-127.
Mizruchi, Mark S., Linda Brewster Stearns, and Christopher Marquis. 2006. "The Conditional Nature of
Embeddedness: A Study of Borrowing by Large U.S. Firms, 1973–1994." American Sociological
Review 71:310-333.
O'Rourke, Dara. 2006. "Multi-stakeholder Regulation: Privatizing or Socializing Global Labor Standards?"
World Development 34:899-918.
Oliver, Pamela E., and Gregory M. Maney. 2000. "Political Processes and Local Newspaper Coverage of
Protest Events: From Selection Bias to Triadic Interactions." The American Journal of Sociology
106:463-505.
Organization for Economic Cooperation and Development (OECD), Working Party of the Trade
Committee. 1999. "Codes of Conduct: An Inventory."
Potoski, Matthew, and Aseem Prakash. 2009. "A Club Theory Approach to Voluntary Programs." in
Voluntary Programs: A Club Theory Perspective, edited by Matthew Potoski and Aseem Prakash.
Cambridge: MIT Press.
Prasad, Monica, Howard Kimeldorf, Rachel Meyer, and Ian Robinson. 2004. "Consumers of the World
Unite: A Market-based Response to Sweatshops." Labor Studies Journal 29:57-80.
Rao, Hayagreeva. 1998. "Caveat Emptor: The Construction of Nonprofit Consumer Watchdog
Organizations." American Journal of Sociology 103:912-61.
Rao, Hayagreeva, Philippe Monin, and Rodolphe Durand. 2003. "Institutional Change in Toque Ville:
Nouvelle Cuisine as an Identity Movement in French Gastronomy." The American Journal of
Sociology 108:795-843.
—. 2005. "Border Crossing: Bricolage and the Erosion of Categorical Boundaries in French Gastronomy."
American Sociological Review 70:968-991.
Rao, Hayagreeva, Calvin Morrill, and Mayer N. Zald. 2000. "Power Plays: How Social Movements and
Collective Action Create New Organizational Forms." Research in Organizational Behavior
22:237-81.
Roberts, Peter W., and Grahame R. Dowling. 2002. "Corporate Reputation and Sustained Superior
Financial Performance." Strategic Management Journal 23:1077-1093.
Rock, Michael T. 2003. "Public Disclosure of the Sweatshop Practices of American Multinational
Garment/Shoe Makers/Retailers: Impacts on Their Stock Prices." Competition & Change 7:23-38.
Ross, Andrew. 1997. No Sweat: Fashion, Free Trade and the Rights of Garment Workers. New York:
Verso.
Sabel, Charles, Dara O'Rourke, and Archon Fung. 2000. "Ratcheting Labor Standards: Regulation for
Continuous Improvement in the Global Workplace." KSG Working Paper No. 00-010
(http://ssrn.com/abstract=253833).
Schneiberg, Marc, Marissa King, and Thomas Smith. 2008. "Social Movements and Organizational Form:
Cooperative Alternatives to Corporations in the American Insurance, Dairy and Grain Industries."
American Sociological Review 73:635-667.
36
Schurman, Rachel. 2004. "Fighting 'Frankenfoods': Industry Opportunity Structures and the Efficacy of
the Anti-Biotech Movement in Western Europe." Social Problems 51:243-68.
Seidman, Gay. 2007. Beyond the Boycott: Labor Rights, Human Rights and Transnational Activism. New
York: Russell Sage Foundation/ASA Rose Series.
Soule, Sarah A. 2009. Contention and Corporate Social Responsibility: Cambridge University Press.
Spar, Debora L. 1998. "The Spotlight and the Bottom Line: How Multinationals Export Human Rights."
Foreign Affairs 77:7-12.
Spar, Debora L., and David B. Yoffie. 2000. "A Race to the Bottom or Governance from the Top?" Pp. 3151 in Coping with Globalization, edited by Aseem Prakash and Jeffrey A. Hart. New York:
Routledge.
Spar, Debra L., and Lane T. LaMure. 2003. "The Power of Activism: Assessing the Impact of NGOs on
Global Business." California Management Review 45:78-101.
Staw, Barry M., and Lisa D. Epstein. 2000. "What Bandwagons Bring: Effects of Popular Management
Techniques on Corporate Performance, Reputation, and CEO Pay." Administrative Science
Quarterly 45:523-556.
United States Department of Labor. 1996. "The Apparel Industry and Codes of Conduct: A Solution to
the International Child Labor Problem?" Washington, DC: U.S. Department of Labor, Bureau of
International Affairs.
Vogel, David. 2005. The Market for Virtue: The Potential and Limits of Corporate Social Responsibility.
New York: Brookings Institution Press.
Weber, Klaus, Kathryn L. Heinze, and Michaela DeSoucey. 2008. "Forage for Thought: Mobilizing Codes
in the Movement for Grass-fed Meat and Dairy Products." Administrative Science Quarterly
53:529-567.
Wetterberg, Anna. 2008. "Codes, Culture, and Coercion: Explaining Adoption of Labor Self-regulation in
the Global Apparel Industry." Paper presented at the annual meeting of the American
Sociological Association, Boston.
World Bank. 2003. "COMPANY CODES OF CONDUCT AND INTERNATIONAL STANDARDS: AN ANALYTICAL
COMPARISON." Washington, DC: The World Bank.
Zajac, Edward J., and James D. Westphal. 2004. "The Social Construction of Market Value:
Institutionalization and Learning Perspectives on Stock Market Reactions." American Sociological
Review 69:433-457.
Zorn, Dirk M. 2004. "Here a Chief, There a Chief: The Rise of the CFO in the American Firm." American
Sociological Review 69:345-364.
Zuckerman, Ezra W. 1999. "The Categorical Imperative: Securities Analysts and the Illegitimacy
Discount." American Journal of Sociology 104:1398-1438.
37